LM3886 availability

Been working doing this since 1969, went through many heavy Argentine meltdowns, (2008 crisis looks like child´s play) meaning Venezuelan style hyperinflation, 400% devaluation in 1 week, 7 presidents in 1 Month (not kidding, in 2002), Military coups, Presidents leaving our "White House" (actually Pink ;) ) in helicopter because they would be murdered in the streets, the whole Monty.
It would be interesting to study how these kinds of issues impact engineering and manufacturing practice throughout the world. In the U.S. it's been far too easy to take supply chain stability for granted. It's a privileged position to be in that comes with potential danger unless you maintain some strategic know-how for when circumstances change.
 
The common theme here is that al these devices have built in speakers. There are a *lot* of them.
Nice catch, hadn´t realized that but of course it´s true.

Or some adventurous person can try the copied ones for quality.
Personally quite the functionalist, so I´d be happy with anybody offering me a "black box", in this case a small black plastic rectangle with 11 pins, couldn´t care less about WHO makes it, even if internal circuit (which I don´t see) is the same or different, as long as reliably meets the specs, can be sourced in quantity and ... ahem!!! .... costs less.

We´ve already talked about my preference of Indian made UR/USHA 2N3055H transistors which I preferred over "official" ON/Motorola or ST ones, go figure, because they met and surpassed my qualifications above.

PRAY
some "USHA" out there manufactures drop-in LM3886 substitutes, even if TI restarts production, just for peace of mind.

Hate being tied to a single supplier.
 
Just a question... was the 3886 ever used in series produced mass market amplifiers, and if so, up to which year, and by which maker?
Mass market means a large scale operation, having at least 5% of the overall market in total of all models.
Not some small workshop operation.

I’m pretty sure I’ve seen it in Cambridge Audio integrated amps of the last ten years , one per channel, and one usually has a hole in it :D
 
The reason why I asked was partly because Chinese copies are already available it seems, a member posted here complaining of fake chips.
He got a pair, one went bang, taking a PCB trace and some other bits.
The other worked, so I asked him how was it, he has not replied yet.
The volumes must be large enough for a proper chip maker to start.
I have already said that some Chinese chip makers put their own marks on knock off chips, and their manpower numbers are startlingly large.
Their own trademark means they stand by their quality, a fake original mark is surely a fake.
A lot of low end and obsolete chips are made by such makers, so the 3886 copy will not be a big deal for makers who make USB controllers, audio chips for computer boards and so on.
So wait till reliable copies are available.
And the new design may be actually improved, a smaller die, and experience with LED driver chips may lead to a better, more compact thermal design, an improvement even!

The other aspect was that the TDA series chip amps were mainstream, I wanted to check that, not being aware of the extent of use of this chip. That is confirmed, the volumes are large, so supplies may be in process.

Usha is a girl's first or proper name, like Maria, or Frances.
It is not an acronym.
 
Last edited:
What’s the matter, MJ15003s just cost too much?
1) sellers asked 3X to 4X what real 3055H cost

2) MJ15003 available here were guaranteed fakes :mad:

3) USHA 2N3055H were *excellent*

Not surprising, they were made by an Indian Military plant for own Military use.

Now you know why :)

Remember I use them by the thousands and in regularly abused products (Guitar amps).
 
Usha is a girl's first or proper name, like Maria, or Frances. It is not an acronym.

So is Svetlana (Russian vacuum tubes)

The LM3886 has been used for years in everything from guitar amps to powered speakers and subwoofers. I had a Brownsville guitar amp with two 12 inch speakers. It had an LM3886 powering each speaker. Plenty WERE used over the years. I suspect that many of the same "boxes with speakers" use class D amp chips today.

Having worked for Motorola for 41 years and being "inside" the semiconductor industry I believe that the decision as to which parts to make going forward is based on a giant spreadsheet with separate pages for each IC process, in a computer somewhere that gets analyzed and discussed quarterly, if not monthly by a team of marketing, financial, and "portfolio optimization" experts.

IC chips are made on different non compatible process lines and each wafer fab can only run one to a few different processes. A line making DSP chips that only see 3 to 5 volts with a "core voltage" in the 1 volt region can not make analog chips that run on tens of volts. The process line capable of making chips like the LM3886 likely also makes linear chips for the automotive market where parts are rated for up to 60 volts. It may also produce discrete semiconductors that run in the same voltage range and can be built on a compatible process.

An IC chip will pass through at least 3 different facilities before being sold. Often these are in different countries. The individual die are produced in a wafer fab. Here lots of chips are built on silicon wafers up to 12 inches in diameter. It takes from 6 to 14 weeks to go from pure silicon to finished wafer.

The wafers are sent to the packaging plant where it is given Go - NoGo testing and cut into individual IC die. The die are then put into whatever packaging (plastic, ceramic, or sometimes sold as bare die), and possible given some more basic testing. Packaging can take a week if you pay for priority, much longer if you are in the normal que.

The packaged parts are then usually sent to another facility where it is given a thorough testing before being put into a saleable container (the tube full of chips). Testing can vary from almost none ( discrete semis and simple logic with mature processes are sample tested) to a full test suite of all parameters and a speed grade out (complex chips like CPU's and memory where the faster chips bring more $$$). This kind of testing takes time, and time is money, so it is the subject of yet another spreadsheet and round of monthly meetings.

The entire semiconductor industry is capacity constrained now, and into the foreseeable future. This puts those "experts" in panic mode meaning there is no clear cut "road map" of future products. There is at least some ranking on that giant spreadsheet, with some clearly green lighted parts. Those would be the big money makers, and parts with sales contracts. These parts are usually in continuous rotation at one or more wafer fab plants.

There are many part in the yellow zone, and these are where most of the "discussion" is focused. Parts in this list may be given short runs on whatever line is available if and when it makes economic sense.

Then there is the orange zone. This may be where the LM3886 lies. Parts here may or may not ever be made again. When a opening becomes available at a fab somewhere, a part from this list may get chosen for a short run. Again it is based on "best economic fit" criteria. This run may be the "lifetime buy" run before product death if that even happens. In times when production is not constrained, the "lifetime buy" status is often based on what's in the pipeline due to a decision to kill the part. This is quite common with small scale chips like voltage regulators of discretes like mosfets.

Will the LM3886 ever see TI production again? I doubt that that decision has been made yet.
 
Having worked for Motorola for 41 years and being "inside" the semiconductor industry I believe that the decision as to which parts to make going forward is based on a giant spreadsheet with separate pages for each IC process, in a computer somewhere that gets analyzed and discussed quarterly, if not monthly by a team of marketing, financial, and "portfolio optimization" experts.

{snipped the rest}

This is the best, most informed post in this entire thread.
 
The Chinese have less variety in chips, and less geographical dispersion.
So they are able to respond faster. Wait and see.
As for the 3886, the order volumes must interest TI enough for the production to be scheduled.

As an aside, China is facing power cuts, and Britain is short on petrol (gasoline).
Another set of causes for delay.
 
This is the best, most informed post in this entire thread.
+1000 :)

I fear Manufacturers think: "Oh, LM3886!!!! ... those gave us good sales and profits, some time ago, but until we solve the 6 Month backlog in MILLIONS of Automotive chips ....forget it"

They said "maybe we´ll make a batch in May next year"? .... thanks for the honesty, I would try to avoid any promised product and stick to what´s currently available from a warehouse anywhere in the World.

That said, maybe a small Factory, too small to even DREAM of getting into the Automotive market, could find a niche for themselves.

Doubt TI would license or outsurce such production, so probably no LM3886 label on chips, not sure if it´s a *brand* or just a "registered type" like, say, 2N3055 or TL072
 
Last edited: